Traffic lights play an import role of directing traffic flows in cities. Currently, the time change periods of traffic lights are substantially predetermined, typically designed based on a principle that opportunities for crossing directions are equal. Further, in a condition where main roads and secondary roads intersect with each other, in view of larger traffic volumes on the main roads, the vehicle passing times of traffic lights are designed based on a principle that the vehicle passing times for the main roads are longer than that for the secondary roads. All the above approaches are based on equal opportunities or probabilities to control vehicle passing times, but do not consider real-time road conditions.
Recently, there are also designs which perform visual recognition on traffic flows and streams of people based on cameras to adjust traffic lights to change passing times at intersections and thereby to effectively improve traffic efficiency. However, the problem of such designs is that the cost for visual recognition is relatively high, and the recognition is not accurate due to the influence of rays of light and occlusions among vehicles.